Respiratory infections including pneumonia and tracheobronchitis remain the leading nosocomial infections in the ICU.1 Ventilator-associated tracheobronchitis (VAT) is manifested by purulent respiratory secretions.1 Differentiating VAT from ventilator-associated pneumonia (VAP) is mainly based on presence/absence of radiographic opacities, but the potential continuum of VAT to VAP is unknown.2,3 The incidence of VAT has been reported to be more common in surgical patients.4 Some reports suggest that patients with VAT also have increased length of ICU stay and a prolonged need for mechanical ventilation.5,6

To gain a better understanding of the molecular mechanisms that underlie the lung phenotypes in VAT vs VAP, we sought to identify and compare their gene expression “signatures” using genome-wide oligonucleotide microarrays. We hypothesized that VAP/VAT would exhibit different gene expression signatures which could help to distinguish between the two conditions and would also shed light on their pathogenesis. We also expected that specific signatures could lead to the identification of new biomarkers to aid in the diagnosis and classification of these diseases.

Eight mechanically ventilated patients were enrolled: five who developed VAT and three VAP. Ventilator-associated infection was diagnosed in all patients after the 4th day. None had received antibiotics before ICU admission or received selective digestive decontamination, and all had started empirical antibiotic therapy when infection was suspected. All but one patient who developed VAP were male. The mean age was 55.2±22.1 years. Subjects were critically ill, with a mean±SD APACHE III score at ICU admission of 75±27. Descriptive data of the cohort are shown in Table 1 (additional material available online Table 1).

The number of samples analyzed in each phase was: in the pre-infection period: 20 samples from 5 VAT patients and 12 samples from 3 VAP patients; in the infection period: 15 samples from 5 VAT patients and 9 samples from 3 VAP patients. To determine the potential influence of quantitative changes of leukocytes subpopulations on our results, we compared the average concentration (cells/mm3) of lymphocytes, neutrophils, eosinophils, basophils and monocytes between VAT and VAP groups both in the pre-infection and infection periods (Table 2) and found no statistical differences.

Microarrays analysis

Comparison of gene expression profiles in the pre-infection period revealed 5595 genes differentially expressed between VAP and VAT (p value <0.01, fold change >2). IPA analysis identified a significant depression of the complement system pathway in the VAP group, affecting to the classical pathway along with the final common pathway (p<0.05) (Fig. 1). Genes found to be differentially expressed in this route were C6, C9, C1QA, C1QB, C1QC, C1S, C3AR1, C4B, C4BPB, CD55, CD59, CR1, and MASP1 (Table 2). In addition, the cAMP and the calcium signalling pathways were also significatively depressed in the VAP group (Fig. 1, Suppl. Table 1) during the pre-infection phase also.

Figure 1.

Complement system gene expression analysis. (a) IPA canonical pathway modelling of the complement system pathway in the pre-infection phase. Those genes with relative lower expression in VAP vs VAT are depicted in green. (b) One-way hierarchical clustering of those genes selected by IPA (red, upregulated; blue, downregulated) in the pre-infection and infection phases in VAP and VAT.

(0.78MB).

On the other hand, comparison of gene expression profiles during the infection period showed 8724 genes differentially expressed between VAP and VAT (p value <0.01, fold change >2). Although not significant differences were found between VAP and VAT for the complement signalling pathway, as showed by IPA, expression of 8 genes participating of this route were significatively depressed in the VAP group compared to the VAT one (Fig. 1). As in the previous phase, IPA evidenced a comparative depression of the cAMP signalling pathway in the VAP group (Suppl. Table 1).

This is the first study to compare gene expression profiles of critically ill patients suffering from VAP and VAT. Gene expression profiles in the pre-infection period identified a significant depression of the complement system pathway in the VAP group compared to the VAT group, affecting to the classical pathway along with the final common pathway. This suggests that intubation by itself is associated with different degrees of immunocompromise that facilitates different infectious complications.

Surveillance cultures every 48–72h in patients with ARDS, using either a protected specimen brush or BAL, have shown that tracheobronchitis can occur before the onset of VAP with tracheal colonization in only 56% of patients and was not predictive of VAP developing. The range of time from colonization to the development of VAP was 2–6 days.10 Conceivably, then, an as yet unidentified intermediate process may allow the bacterial burden to develop VAT before VAP, and this intermediate period may constitute an excellent time window in which to halt the progression.

Other data in mechanically ventilated patients suggest a similar dichotomy between upper and lower airway colonization patterns. As the upper airways have contact with pathogenic microbes, immune recognition principles have to be tightly controlled.

In our study we mainly observe a suppression of genes involved in the classical pathway and converge for common pathway of complement activation. Complement activation is a key component for maintaining internal inflammatory homeostasis. The complement system is a set of more than thirty proteins which serves as an important effector arm in the immune defense and contributes to the first-line host defense.11,12 It comprises an essential link between the innate and adaptive immune system,13,14 as its proteins identify invading bacteria and label them for phagocytosis or lysis.

Monocytes actively produce proteins of the complement system. In consequence, altered numbers of monocytes in peripheral blood could contribute to explain the observed differences in the gene expression profiles between groups. To determine the potential influence of quantitative changes of leukocytes subpopulations on our results, we compared the average concentration of lymphocytes, neutrophils, eosinophils, basophils and monocytes between VAT and VAP groups both in the pre-infection and infection periods and revealed no significant differences between groups. This supports that relative depression in the expression values of the genes associated to complement function in the VAP group compared to VAT in the pre-infection period should be potentially due to functional changes induced by the disease, and not to quantitative changes in leukocyte subpopulations.

Comparing the gene expression profiles of VAP and VAT, the cAMP signalling pathways were significantly depressed in the VAP group. During the development of the immune systems, cAMP regulates cell proliferation, differentiation and apoptosis.15 Recent studies have demonstrated that cAMP is necessary for T cell activation.16 In contrast, other studies have shown it to be a potent immunosuppressant: in murine dendritic cell (DC) models, other authors17 have demonstrated that increased levels of cAMP in DC down-regulated the stimulatory molecules CD80 and CD86. Another pathway with differential gene expression was the calcium signalling pathway. Calcium is involved many biological and immune system functions such as T lymphocyte activation,18 the immunological synapse (IS) between T cells and antigen-presenting cells19 and phagocytosis.20 However, the depression of the cAMP and Ca pathway genes is difficult to interpret as almost every cell uses these pathways at the initiation of different signalling pathways and this effect should be elucidated in future studies.

In our study, patients who developed VAP had lower gene expression in the complement signalling pathways. In recent years, immodulatory therapies in sepsis were driven by the assumption that the adverse outcomes were related to an overly exuberant inflammatory response. However, recent evidence on cytokine production in response to bacterial antigens, lymphocyte proliferation in response to recall antigens, and new antigen presentations as markers of the immune function has been published. Immunoparalysis should be taken into account as a continuum along which patients with VAP presented a less compartimentalised infection.

The present study has some strengths and weaknesses. First, a control population of patients matched by severity and diagnoses who underwent ventilation for equivalent times but did not develop VAP or VAT would be advisable. Unfortunately, lack of financial support precludes microarray analysis of controls. Therefore, our findings constitute a pilot study and needs to be validated before generalization, because selection bias cannot be excluded.

Second, VAP does not have gold standard. Indeed, there was a possible overlap in clinical signs between VAT and VAP, and ICU chest X-rays have limitations in sensitivity to identify opacities. Indeed, as Kollef21 suggested, episodes of VAT might constitute only airways colonization, or perhaps small pneumonia in some patients. CT scan was not implemented for diagnosis because this technique is expensive and is not part of standard care; in addition, transfers to CT scan in intubated patients present safety concerns. Moreover, other opacities such as atelectasis or ARDS cannot be ruled out.22 Indeed, Craven et al.22 have acknowledged this potential bias. The third major limitation of our study is that we only followed patients until infection onset. Cobb et al.23,24 reported that the onset of an infection-specific transcriptional program may precede the clinical diagnosis in patients who developed VAP based on the Riboleukogram. The main aim of that study was to determine whether VAP/VAT exhibits different gene expression signatures that might explain divergent pathways leading to VAT/VAP as a continuum between bronchitis and pneumonia in mechanically ventilated patients. Fourth, our findings cannot be generalized to pneumonias developing after one week of ventilation, but 80% of episodes of VAP present during the first 10 days.25,26 Finally, the sample size is relatively small and gene expression profiles between Gram-positive and Gram-negative pathogens could not be determined, but the number of samples obtained was sufficient to generate statistically significant differences with a large number of genes being functionally involved in the immune response. This is the first study to investigate the biological plausibility between VAT and VAP.

In summary, this study suggests that intubated patients with VAP show a relative depression of the genes involved in the complement system, cAMP, and calcium signalling pathways which all play a key role in the immune response to bacteria. The relative depression of these routes may lead to an impaired immunocompetence status, and contributes to explain why patients with VAP exhibit more inflammation and worse outcomes than those with VAT. Therefore, therapy of VAP should be more aggressive than VAT. Our findings suggest that VAP does not have an abrupt onset and risk of VAP might be anticipated by identification of the signature that precedes pneumonia, facilitating pre-emptive therapy. Moreover, this translational study offers insight to improve the understanding of pathogenesis of VAP vs VAT, with implications on prevention. Indeed, early development of immunocompromise in the post-intubation period suggests that not all episodes of VAP are preventable and “pneumonia zero” is not a realistic objective.

The authors declare that they have no conflict of interest.